Hugging Face's logo

tyzhu
/
lmind_hotpot_train8000_eval7405_v1_qa_3e-5_lora2

Safetensors
Generated from Trainer
Eval Results
Model card Files Community
lmind_hotpot_train8000_eval7405_v1_qa_3e-5_lora2 / README.md
tyzhu's picture
tyzhu
End of training
d5dd343 verified
preview code
|
raw
history blame
3.17 kB
metadata
license: llama2
base_model: meta-llama/Llama-2-7b-hf
tags:
  - generated_from_trainer
datasets:
  - tyzhu/lmind_hotpot_train8000_eval7405_v1_qa
metrics:
  - accuracy
model-index:
  - name: lmind_hotpot_train8000_eval7405_v1_qa_3e-5_lora2
    results:
      - task:
          name: Causal Language Modeling
          type: text-generation
        dataset:
          name: tyzhu/lmind_hotpot_train8000_eval7405_v1_qa
          type: tyzhu/lmind_hotpot_train8000_eval7405_v1_qa
        metrics:
          - name: Accuracy
            type: accuracy
            value: 0.5824050632911393

lmind_hotpot_train8000_eval7405_v1_qa_3e-5_lora2

This model is a fine-tuned version of meta-llama/Llama-2-7b-hf on the tyzhu/lmind_hotpot_train8000_eval7405_v1_qa dataset. It achieves the following results on the evaluation set:

  • Loss: 2.9797
  • Accuracy: 0.5824

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

  • learning_rate: 3e-05
  • train_batch_size: 2
  • eval_batch_size: 2
  • seed: 42
  • distributed_type: multi-GPU
  • num_devices: 4
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 32
  • total_eval_batch_size: 8
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: constant
  • lr_scheduler_warmup_ratio: 0.05
  • num_epochs: 20.0

Training results

Training Loss Epoch Step Validation Loss Accuracy
1.8255 1.0 250 1.8392 0.6054
1.7368 2.0 500 1.8111 0.6078
1.6689 3.0 750 1.8103 0.6075
1.5555 4.0 1000 1.8414 0.6067
1.4559 5.0 1250 1.8992 0.6038
1.3514 6.0 1500 1.9584 0.6018
1.2491 7.0 1750 2.0300 0.6000
1.1749 8.0 2000 2.1051 0.5982
1.0769 9.0 2250 2.1948 0.5954
1.0134 10.0 2500 2.2515 0.5943
0.9209 11.0 2750 2.3421 0.5921
0.8636 12.0 3000 2.4443 0.5905
0.7866 13.0 3250 2.5574 0.588
0.7448 14.0 3500 2.5800 0.5867
0.6709 15.0 3750 2.6912 0.5846
0.6439 16.0 4000 2.7546 0.5853
0.5869 17.0 4250 2.7997 0.5831
0.5596 18.0 4500 2.8435 0.5833
0.5205 19.0 4750 2.9510 0.5833
0.5045 20.0 5000 2.9797 0.5824

Framework versions

  • Transformers 4.34.0
  • Pytorch 2.1.0+cu121
  • Datasets 2.18.0
  • Tokenizers 0.14.1